The Forgotten Middle: Many Middle-Income Seniors Will Have Insufficient Resources For Housing And Health Care
- Caroline F. Pearson ([email protected]) is a senior vice president at NORC at the University of Chicago in Illinois.
- Charlene C. Quinn is an associate professor in the Department of Epidemiology and Public Health at the University of Maryland School of Medicine, in Baltimore.
- Sai Loganathan is a senior health economist at NORC at the University of Chicago in Bethesda, Maryland.
- A. Rupa Datta is a vice president at NORC at the University of Chicago in Illinois.
- Beth Burnham Mace is chief economist at the National Investment Center for Seniors Housing and Care, in Annapolis, Maryland.
- David C. Grabowski is a professor in the Department of Health Care Policy, Harvard Medical School, in Boston, Massachusetts.
As people age and require more assistance with daily living and health needs, a range of housing and care options is available. Over the past four decades the market for seniors housing and care—including assisted living and independent living communities—has greatly expanded to accommodate people with more complex needs. These settings provide housing in a community environment that often includes personal care assistance services. Unfortunately, these settings are often out of the financial reach of many of this country’s eight million middle-income seniors (those ages seventy-five and older). The private seniors housing industry has generally focused on higher-income people instead. We project that by 2029 there will be 14.4 million middle-income seniors, 60 percent of whom will have mobility limitations and 20 percent of whom will have high health care and functional needs. While many of these seniors will likely need the level of care provided in seniors housing, we project that 54 percent of seniors will not have sufficient financial resources to pay for it. This gap suggests a role for public policy and the private sector in meeting future long-term care and housing needs for middle-income seniors.
Health problems, cognitive decline, and mobility limitations that become more common with aging often result in older people struggling to live independently. As care needs increase, people pursue a range of solutions—including increased reliance on family caregivers, in-home support services, or both and moving out of their homes into alternative housing arrangements. A number of housing options exist, including private independent living and assisted living communities for those with a range of care needs and nursing homes when high-intensity care is required. Housing options available to older people often depend on the financial resources available to them.
Over the past forty years the housing market for seniors (people ages seventy-five and older), particularly assisted living and independent living communities, has experienced tremendous growth and evolution.1 Seniors housing provides residence and care to about two million older adults, roughly evenly split between independent and assisted living.2,3 Increasingly, people living in these communities have high rates of chronic illness, functional dependence, and medical complexity.2,4 In the context of this higher acuity, seniors generally have a desire to live and receive care in the least institutional and most homelike setting possible.5 The emergence of seniors housing has been influenced by consumer preference for the high level of socialization and autonomy found in these communities, coupled with access to personal care assistance services. Indeed, private-paying residents who were previously cared for in nursing homes now often receive services in assisted living.5
Seniors housing operators and investors have largely focused on the upper end of the income distribution. For lower-income people, state and local programs provide housing and care services via means-tested programs such as Medicaid. For higher-income people, a large private-pay assisted or independent living sector has emerged to meet the demand for housing and care services.3,6
Although some middle-income people are living in seniors housing, the industry has not primarily focused on this cohort. This income group is generally too wealthy to qualify for public means-tested programs, yet not wealthy enough to pay the costs at many seniors housing communities for a sustained period of time. Medicare does not cover seniors housing or long-term care services. About 0.7 million elderly people who would not normally qualify for public programs on the basis of their income and assets ultimately become eligible for Medicaid coverage in the community or nursing home by spending down on medical costs and a mix of paid and informal caregivers.7
With the aging baby-boom generation, the US will experience a significant increase in the number of middle-income seniors ages seventy-five and older by 2029. This group will face a very different set of challenges relative to today’s middle-income seniors (defined using 2014 data). Future seniors have lower overall savings and are less likely to have pensions, as defined-contribution retirement plans have grown. This trend may increase pressure on the already decreasing number of familial caregivers per senior.8 Consequently, Medicaid, the dominant payer of long-term care in the US, could see ballooning demand as the number of seniors in need of such care increases. A key policy question is how future middle-income seniors, who do not qualify for Medicaid and may have fewer family caregivers, will access housing and care services.
The purpose of this study was to analyze the growth from 2014 to 2029 in the number of middle-income older adults, focusing on seniors who will be ages seventy-five and older at that time. Who are they? What will their care needs be? Will they have sufficient financial resources to access seniors housing? Previous research has projected the future wealth of older adults.9 To our knowledge, this is the first study to project housing and functional health care needs of seniors by income group. This study creates an opportunity for policy makers to break down silos that heretofore have separately considered housing and health care needs.
Study Data And Methods
This analysis forecast the size of the US population ages seventy-five and older in 2029 (“tomorrow’s seniors”) and estimated their demographic characteristics, health, cognitive and functional status, and financial resources. The analytic model was developed in three stages: the construction of an individual financial resource measure that included income, annuitized assets, and annuitized housing equity; forecasting of the size and demographic characteristics of the senior population in 2029; and estimation of per capita financial resources and selected health and functional characteristics of the forecasted senior population. More details about the methodology are in online appendix E.10
This analysis used data from the Health and Retirement Study (HRS), a nationally representative, longitudinal survey of people ages fifty and older.11 The HRS—sponsored by the National Institute on Aging and conducted by the University of Michigan—includes information on health and functional status, financial resources, and sociodemographic characteristics. HRS cross-sectional sampling weights were used to produce nationally representative estimates of today’s (2014) senior population and people projected to be seniors by 2029.
A distinguishing feature of the analytic approach is the measurement of personal rather than household financial resources. People’s finances can be linked to their health and functional status to understand who is more likely to need care and what resources they will have to fund it. Individual-level finances better reflect the transitions that seniors experience after the loss of a spouse or partner.
For the financial resources measure, income and nonhousing assets were annuitized and reported at the individual level using published models12–14 and Social Security Administration life expectancy calculations. Housing equity (net of mortgage debt and home loans) was annuitized but held separately from other assets.14
To build a regression model to predict the number of years of remaining life, people in the 1994 HRS were followed until 2014. We modeled life expectancy using a multivariate tobit model to control for demographic, socioeconomic, education, health, and functional factors.9 This model was applied to the 2014 HRS data to predict the life expectancy of each person. Those predicted to be alive in 2029 constitute the forecasted sample of the senior population in 2029.
Because demographics and certain socioeconomic characteristics (for example, education) can be considered time-invariant for people ages fifty-five and older, we applied population proportions of these characteristics to the forecasted 2029 senior population. We assumed the marital status of people would be unchanged from 2014 to 2029, unless they were forecasted to lose their spouse or partner during that time frame (new marriages and divorces are not included in the forecast).
Health, cognitive, and functional status of the forecasted senior population in 2029 were estimated based on the age group, income cohort, sex, and race/ethnicity of people forecasted to survive to 2029. Cognitive impairment was defined using a published composite measure,15 and mobility limitations were defined using the HRS mobility index. Seniors with high needs were defined as those who had three or more chronic conditions and a limitation in at least one activity of daily living (ADL).16
To predict the annuitized individual financial resources (income) of seniors in 2029, we began with the cohort’s actual income as of 2014 and calculated the expected change between 2014 and 2029 using the inflation-adjusted change from 1998 to 2014 by financial component (income, annuitized assets, and annuitized housing equity). Income cohort– and age group–specific trend factors were applied to people in the 2014 sample who were expected to be seniors in 2029. Because the financial resource calculation converted household assets to individual resources, the model included a transition of people from being married (sharing assets) to being widowed (solely owning assets).
Our analysis focused on middle-income seniors from the perspective of seniors housing affordability. We segmented the 2014 senior population into three groups according to their annuitized individual financial resources (excluding housing equity), defining the middle-income cohort as those in the forty-first to the eightieth percentile of the individual financial resource distribution of today’s (2014) seniors. Our definition of the middle-income cohort was motivated by its relevance to assessing the affordability of seniors housing and the feasibility of conducting the analysis within the sample size constraints of the HRS. We set a conservative minimum threshold at which people are unlikely to qualify for Medicaid. Our upper threshold was set based on the annual costs of seniors housing and the sample size of the high-income group. In 2029, for people ages 75–84, that middle-income definition corresponds to annuitized financial resources of $25,001–$74,298 in 2014 dollars. For those ages 85 and older, middle income is $24,450–$95,051. To make direct comparisons between the middle-income cohorts of 2014 and 2029, we projected the 2029 middle-income cohort based on 2014 dollar thresholds. (See appendix A for the thresholds for low, middle, and high incomes.)10
Our study had several limitations. First, this analysis asked what the need for and affordability of seniors housing would look like in 2029 if recent health, demographic, and financial trends continued. To the extent that health conditions are becoming more or less common within demographic subgroups, our study may have misestimated the number of tomorrow’s seniors who will be affected by them.
Second, we included a standard assumption that seniors will have $5,000 in average annual out-of-pocket medical spending.17 This represents a conservative assumption since $5,000 is at the very low end of average seniors’ medical expenses, especially for those older than eighty-five with multiple chronic conditions and functional limitations.17 Seniors’ out-of-pocket health care spending will vary widely based on their health status and use of medical services, whether they are enrolled in fee-for-service Medicare or Medicare Advantage, and whether they purchase supplemental coverage. Trends in health care costs or Medicare policy changes could significantly affect this assumption.
Third, although we used the actual 2014 financial resources of the people who will be seniors in 2029, we assumed that the rate of change in income, assets, and housing equity from 2014 to 2029 would be the same as the historical trend from 1998 to 2014. Differences in economic performance over the next ten years could affect our overall and subgroup projections.
Finally, our estimates for the costs of seniors housing represent national average rates with limited levels of services. In fact, rates vary widely by geography, and costs increase with higher levels of service. Seniors with mobility, cognitive, or other functional limitations are likely to need additional care services not included in the estimated costs of seniors housing. We also assumed that future seniors housing options will exist as they do today, with costs growing at the rate of inflation. This is a conservative assumption, which may underestimate seniors housing costs if increased demand is not met with a proportionate increase in supply.
Number Of Middle-Income Seniors
As baby boomers age, the overall senior population will grow dramatically (exhibit 1). Results show a dramatic reduction in the percentage of low-income seniors and a rapid rise in the share of high-income seniors over the period 2014–29. Meanwhile, middle-income seniors—who are the focus of this study because of the care options available to them—are projected to nearly double, from 7.9 million to 14.4 million. More of the growth by 2029 will be concentrated in seniors ages 75–84, as the oldest boomer will be only eighty-three in that year. Middle-income seniors will grow in absolute numbers and will also account for a larger share of the total senior population, increasing from 40 percent of the senior population in 2014 to 43 percent in 2029. Our projections show that the number of middle-income seniors ages 75–84 will increase from 5.57 million in 2014 to 10.81 million in 2029.
Middle-income seniors in 2029 will also differ demographically from today’s seniors (exhibit 2). The growth in the number of women will outpace that of men, with women constituting 58.3 percent of middle-income seniors in 2029 compared to 56.0 percent in 2014. Future middle-income seniors will also be more racially diverse than today’s seniors, with increases in the shares of Hispanics (from 2.4 percent to 6.3 percent) and non-Hispanic blacks (from 5.0 percent to 6.7 percent). Lower growth rates for non-Hispanic whites will result in their share shrinking from 90.9 percent to 83.8 percent. Finally, this population will also become more educated in the next decade, as the number of people who did not graduate from high school will decline and those with a four-year college degree and above will increase by 59.6 percent. This change directly contributes to a rightward shift in the forecasted income distribution of future seniors because higher education levels result in fewer seniors living in poverty.
|Millions of middle-income seniors|
|Ages in 2014 (years)||Ages in 2029 (years)||Percent of total (ages 75 and older)|
|75–84||85 and older||75 and older||75–84||85 and older||75 and older||2014||2029|
|Less than high school graduate||0.62||0.34||0.96||0.25||0.23||0.48||12.1||3.3|
|High school graduate||2.30||0.98||3.28||2.80||1.42||4.22||41.3||29.4|
|College and above||1.41||0.50||1.91||4.36||1.15||5.51||24.1||38.4|
|Home, no assistance needed||4.00||1.11||5.11||— a||— a||— a||64.4||— a|
|Home, assistance needed||1.00||0.65||1.68||— a||— a||— a||21.2||— a|
|Outside the home||0.52||0.59||1.11||— a||— a||— a||14.0||— a|
One important demographic shift that could significantly affect middle-income seniors’ long-term housing and care needs is the future availability of caregivers.8 Currently, only 14 percent (1.11 million) of middle-income seniors reside outside the home in a community setting (assisted or independent living) or nursing facility. Among those who remain at home, one in four (1.68 million) require assistance that is often provided by a family member (see appendix B).10
Seniors who have mobility and cognitive limitations and those with multiple chronic conditions are more likely to require additional care and support with basic needs for a longer period. A substantial portion of tomorrow’s middle-income seniors are projected to have health and mobility limitations (exhibit 3). Notably, we did not assume any change in the prevalence of these conditions among our income-age subgroups from 2014. Some conditions, such as obesity, are likely to increase, while other disabling conditions may become less common or be more effectively managed, which would affect these projections. We estimated that 20 percent of tomorrow’s middle-income seniors will fall into the “high needs” group (people with three or more chronic conditions and one or more limitations in ADLs). Our projections show that 60 percent of tomorrow’s middle-income seniors will have mobility limitations that may prevent them from remaining independent and in their homes. These mobility limitations tend to increase with age, with 73 percent of middle-income seniors ages eighty-five and older projected to have a mobility limitation. Lastly, 6 percent of middle-income seniors ages 75–84 and 15 percent of those ages 85 and older are projected to have cognitive impairments.
|Ages 75–84||Ages 85 and older||Ages 75 and older|
|Millions||% of age group||Millions||% of age group||Millions||% of age group|
|All middle-income seniors||10.81||100.0||3.54||100.0||14.35||100.0|
|3 or more chronic conditions||6.97||64.5||2.64||74.6||9.61||67.0|
|3 or more ADLs||0.64||5.9||0.60||16.9||1.24||8.6|
|Mobility limitations and cognitive impairment||0.41||4.0||0.43||12.0||0.84||5.9|
Although health and mobility needs necessitate additional care and support, many of tomorrow’s middle-income seniors might not have the financial resources required to pay for private seniors housing options as they exist today. In 2029, 11.6 million (81 percent) middle-income seniors without equity in housing will have annual income and annuitized assets (henceforth, “income”) of $60,000 or less (exhibit 4). This will be below the projected average annual cost of $62,000 for assisted living rent and estimated medical out-of-pocket spending. Medical costs and rent will vary by individual and geographic region. Even assuming that seniors draw from their housing equity in addition to their income, 7.8 million (54 percent) middle-income seniors in 2029 will have annual financial resources of $60,000 or less. This gap in financial resources is particularly pronounced for seniors ages 75–84, 6.4 million (59 percent) of whom will have annual financial resources with housing equity of $60,000 or less and the average assisted living rent plus medical expenses (see appendix C).10 Among middle-income seniors ages 85 and older, 1.4 million (39 percent) will have annual financial resources below this level when housing equity is included. Older people can have greater financial resources because they are more likely to be widowed and therefore individually own assets that they previously held jointly with their spouse (see appendix D).10
This study is unique in its emphasis on middle-income seniors, a group often overlooked by both policy makers and the investors in and operators of private seniors housing. Our definition of middle income was motivated in part by the seniors housing options that exist in the market today. We conservatively selected a definition that identified seniors who would be unlikely to qualify for Medicaid long-term care. On the upper end, we excluded higher-income seniors who could access the current seniors housing product offerings. A positive result is a shift in the income distribution that reduces the number of poor seniors, many of whom would have relied on Medicaid, and increases the number of high-income seniors, many of whom will have the financial resources to choose their preferred options for housing and care as they age. However, the middle-income cohort, which has more limited care options, is also growing in size and proportion.
As people age, their choices about where to live may be informed by their health, the availability of caregivers, their financial resources, and their desire for socialization. As our results show, a large and growing cohort of middle-income seniors will have a gap between their housing and care needs and the financial resources to meet those needs, given today’s seniors housing options. As they age, many middle-income seniors may suffer from declines in health and mobility that prevent them from living independently. Limitations or anticipated limitations in mobility are common reasons why people move out of their homes and into alternative housing arrangements,18 and 60 percent of tomorrow’s middle-income seniors will have some mobility limitations. Furthermore, one in five of these seniors will be high needs (having three or more chronic conditions and one or more limitations in ADLs). These seniors are unlikely to be able to remain in their homes without meaningful support from family or paid caregivers. Our ten-year model ended before the large cohort of seniors turns eighty-five, when rates of health conditions increase rapidly. Thus, the number of seniors needing additional care is likely to grow dramatically after 2029.
Spouses and middle-age daughters constitute the bulk of family caregivers. However, the availability of these caregivers has declined as a result of changing marriage patterns, lower birth rates, and where adult children live and work.8 Despite this, the total amount of care provided by familial caregivers has remained steady from 1989 to 2012,8 though it is unclear whether this trend will hold as the senior population continues to grow. Future seniors with care needs who do not have access to family support will likely seek to shift into seniors housing or seek paid care from professionals or unpaid caregivers.
Unfortunately, our study revealed that most of tomorrow’s middle-income seniors will lack the financial resources required to pay for private seniors housing, regardless of their preferences. Even if we assume that seniors devote 100 percent of their annual income to seniors housing—setting aside any personal expenses—only 19 percent of middle-income seniors will have financial resources that exceed today’s costs of assisted living. Many seniors treat housing equity differently from other financial resources and attempt to liquidate other income and assets before liquidating the equity. Such housing equity may be the family home that some older adults keep as a nest egg to protect against future, unexpected financial hardship or wish to preserve for their children. However, if we assume that middle-income seniors do draw down housing equity, 54 percent (7.8 million) will still lack the resources to pay for seniors housing at today’s costs.
This confluence of factors creates a significant unmet future need, which demands new housing and care solutions to support the emerging generation of America’s seniors. Creating and financing those solutions will require innovation from public and private stakeholders to bring more affordable seniors housing options to the market and to enable people at all income levels to access the care they need and want as they age.
Considerations For The Future
Alternative solutions are needed to serve the millions of seniors who cannot remain in their homes and who will lack the financial resources for private-pay seniors housing. Among these solutions is finding less costly ways to provide housing and care.
From the private-sector perspective, options may include lowering investment return expectations by charging less rent and reducing profit margins; subsidizing lower-income residents with higher-paying residents, as in mixed-income communities; or offering more basic and less expensive housing service products. Technology may be another solution to reduce operating costs, increase staff efficiency, or make residents more self-sufficient. Seniors housing could also work to more formally involve family caregivers, outside volunteers, and healthier residents in a structured way to offset staffing costs, as occurs in some other countries. Finally, à la carte pricing models that break out service and care expenses from housing create more flexibility for some residents.
Although some seniors housing operators are experimenting with options today, they are not widespread. One way to stimulate such private-sector innovation could be to offer tax incentives for developers and operators of seniors housing to serve middle-income seniors. As the opportunity to serve this growing cohort becomes more recognized, we expect creative entrepreneurs to pursue other yet-to-be-imagined solutions.
From a public program perspective, housing and health care policies could be considered. For example, within federal housing policy, such as that related to low-income tax credits and other programs, eligibility limits could be raised to include more middle-income seniors. On the consumer side, subsidies or voucher programs could be expanded to allow more seniors to access seniors housing or new incentives created to encourage long-term care financial planning.
Another possibility is for housing communities with sufficient capacity to establish their own Medicare Advantage plans. For example, one plan offers on-site medical services delivered by employed physicians.4 Other innovative plans being developed include provider consortiums that jointly own Medicare Advantage plans to capture covered lives on a more cost-effective basis.19
Beginning in 2019, Medicare Advantage plans may offer supplemental benefits that cover nonmedical services, including in-home modifications. However, such benefits are optional for Medicare Advantage plans offered by companies that believe the coverage will be cost-effective; they are not available under fee-for-service Medicare. Reimagining the Medicare benefit to provide broader access to supportive services could be one way to address the impending unmet needs. Alternatively, lawmakers could consider a new benefit that explicitly funds long-term care (for example, a Medicare “Part E” that shifts funds from Medicare Part A acute care).
From a Medicaid policy perspective, policy makers could consider changes to Medicaid long-term care benefits. The current program requires people to impoverish themselves (“spend down”) to qualify for coverage. Although Medicaid is the payer of last resort, even high-income people can eventually receive sizable Medicaid coverage by living longer and having higher medical needs in old age.20 However, by broadening eligibility and expanding coverage to home and community-based services for beneficiaries with higher incomes and less acute health needs, the program may be able to forestall health and functional deterioration and keep seniors in noninstitutional settings longer, when preferred.
Today, roughly one-fifth of assisted living residents have some of their care covered by Medicaid, though many programs have waiting lists and other coverage limits.21 By statute, Medicaid can pay only for care services and is not allowed to cover the costs of housing, except for institutions such as nursing homes and hospitals. State Medicaid programs could take several measures to make these housing costs more affordable, such as limiting the amount seniors housing communities can charge Medicaid beneficiaries, allowing residents to retain greater income to pay for room and board while still qualifying for Medicaid, or allowing supplementation by families or trusts.22
Any policy solution should recognize the full range of services that seniors may need as they age.
Importantly, any policy solution should recognize the full range of services that seniors may need as they age. Failure to provide in-home supportive services may lead to a premature loss of function and avoidable institutionalization for some people. Meanwhile, remaining in one’s home alone might not produce the best outcomes for all people. Social isolation and loneliness have been documented to create real and costly declines in health status.23 Future initiatives to expand seniors housing options and affordability should recognize a range of possible living arrangements, including in-home and community-based options.
Our forecast of the 2029 middle-income senior population reveals a growing need for seniors housing options. As baby boomers age, middle-income seniors will become a larger and more diverse group, many of whom will have health, cognitive, or mobility limitations that make it difficult for them to live independently. A reduction in the number of family caregivers places greater burden on those remaining, and some seniors may not have unpaid care options available. Unfortunately, most middle-income seniors will not have the financial resources required to pay for private seniors housing options, as they exist today. Innovations from policy makers and private-sector housing operators and their investors will be required to serve this middle-income cohort.
The National Investment Center for Seniors Housing & Care (NIC) provided a grant to NORC at the University of Chicago to fund this research. Charlene Quinn is a paid research and policy adviser to NIC. David Grabowski serves as a paid consultant to Vivacitas and CareLinx. He also serves on the Scientific Advisory Committee of NaviHealth. The authors acknowledge and thank Robert G. Kramer, founder of and strategic adviser at NIC, for his insights, contributions, and guidance; and Lana Peck, senior principal, and Chuck Harry, chief of Research & Analytics, both at NIC, for their help in providing research and support. [Published online April 24, 2019.]
- 1 . Historical evolution of assisted living in the United States, 1979 to the present. Gerontologist. 2007;47(Spec No 3):8– 22. Crossref, Medline, Google Scholar
- 2 . The residential continuum from home to nursing home: size, characteristics, and unmet needs of older adults. J Gerontol B Psychol Sci Soc Sci. 2014;69(7, Suppl 1):S42– 50. Crossref, Medline, Google Scholar
- 3 . Sizing up the market for assisted living. Health Aff (Millwood). 2010;29(1):35– 43. Go to the article, Google Scholar
- 4 . Integrated payment and delivery models offer opportunities and challenges for residential care facilities. Health Aff (Millwood). 2015;34(10):1650– 6. Go to the article, Google Scholar
- 5 . Assisted living expansion and the market for nursing home care. Health Serv Res. 2012;47(6):2296– 315. Crossref, Medline, Google Scholar
- 6 . The asset and income profiles of residents in seniors housing and care communities: what can be learned from existing data sets. Res Aging. 2013;35(1):50– 77. Crossref, Google Scholar
- 7 . The Medicaid medically needy program: spending and enrollment update [Internet]. Washington (DC): Kaiser Commission on Medicaid and the Uninsured; 2012 Dec [cited
2019 Mar 7]. (Issue Paper No. 4096). Available from: https://kaiserfamilyfoundation.files.wordpress.com/2013/01/4096.pdf Google Scholar
- 8 . Trends in informal care for disabled older Americans, 1982–2012. Gerontologist. 2018;58(5):863– 71. Crossref, Medline, Google Scholar
- 9 . The dynamic simulation of income model (DYNASIM): an overview [Internet]. Washington (DC): Urban Institute; 2015 Sep [cited
2019 Mar 7]. Available from: https://www.urban.org/sites/default/files/publication/67366/2000391-The-Dynamic-Simulation-of-Income-Model-DYNASIM-%20An-Overview.pdf Google Scholar
- 10 To access the appendix, click on the Details tab of the article online.
- 11 Health and Retirement Study Survey Research Center. Health and Retirement Study, 1994, 1998, and 2014 core and cross-wave tracker public use datasets [Internet]. Ann Arbor (MI): University of Michigan; 2017 Dec [cited
2019 Mar 23]. Available [registration required] from: http://hrsonline.isr.umich.edu/index.php?p=avail&_ga=2.2281264.1229203199.1552332403-793281653.1551127024 Google Scholar
- 12 . Joint life annuities and annuity demand by married couples [Internet]. Cambridge (MA): National Bureau of Economic Research; 1999 Jun [cited
2019 Mar 7]. (NBER Working Paper No. 7199). Available from: https://www.nber.org/papers/w7199.pdf Google Scholar
- 13 . The trajectory of wealth in retirement. J Public Econ. 2009;93(1–2):191– 208. Crossref, Google Scholar
- 14 . The composition and drawdown of wealth in retirement. J Econ Perspect. 2011;25(4):95– 118. Crossref, Medline, Google Scholar
- 15 . Measures of cognitive functioning in the AHEAD Study. J Gerontol B Psychol Sci Soc Sci. 1997;52(Spec No):37– 48. Crossref, Medline, Google Scholar
- 16 High-need, high-cost patients: who are they and how do they use health care? A population-based comparison of demographics, health care use, and expenditures [Internet]. New York (NY): Commonwealth Fund; 2016 Aug 29 [cited
2019 Mar 7]. Available from: https://www.commonwealthfund.org/publications/issue-briefs/2016/aug/high-need-high-cost-patients-who-are-they-and-how-do-they-use Google Scholar
- 17 . Medicare beneficiaries’ out-of-pocket health care spending as a share of income now and projections for the future [Internet]. San Francisco (CA): Henry J. Kaiser Family Foundation; 2018 Jan [cited
2019 Mar 7]. Available from: http://files.kff.org/attachment/Report-Medicare-Beneficiaries-Out-of-Pocket-Health-Care-Spending-as-a-Share-of-Income-Now-and-Projections-for-the-Future Google Scholar
- 18 . Functioning, forgetting, or failing health: which factors are associated with a community-based move among older adults? J Gerontol B Psychol Sci Soc Sci. 2016;71(6):1120– 30. Crossref, Medline, Google Scholar
- 19 . New senior living consortium to launch Medicare Advantage network. Senior Housing News [serial on the Internet]. 2019 Feb 11 [cited
2019 Mar 7]. Available from: https://seniorhousingnews.com/2019/02/11/new-senior-living-consortium-to-launch-medicare-advantage-network/ Google Scholar
- 20 . Medicaid insurance in old age [Internet]. Cambridge (MA): National Bureau of Economic Research; 2013 Jun [cited
2019 Mar 7]. (NBER Working Paper No. 19151). Available from: https://www.nber.org/papers/w19151.pdf Google Scholar
- 21 . Medicaid in residential care [Internet]. Washington (DC): Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation; 2013 Sep [cited
2019 Mar 7]. Available from: https://aspe.hhs.gov/system/files/pdf/177331/FacBenDif.pdf Google Scholar
- 22 . Compendium of residential care and assisted living regulations and policy: 2015 edition [Internet]. Washington (DC): Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation; 2015 Jun [cited
2019 Mar 7]. Available from: https://aspe.hhs.gov/system/files/pdf/110391/15alcom.pdf Google Scholar
- 23 Perceived social isolation and outcomes in patients with heart failure. J Am Heart Assoc. 2018;7(11):e008069. Crossref, Medline, Google Scholar